It is standard in multiagent settings to assume that agents will adopt Nash equilibrium strategies. However, studies in experimental economics demonstrate that Nash equilibrium is a poor description of human players ’ actual behaviour. In this study, we consider a wide range of widely-studied models from behavioural game theory. For what we believe is the first time, we evaluate each of these models in a meta-analysis, taking as our data set large-scale and publicly-available experimental data from the literature. We then propose a modified model that we believe is more suitable for practical prediction of human behaviour. ii Table of Contents Abstract................................... ii

Current implementations of agent-based online auction systems only support simple predefined bidding strategies for bidding agents. In this paper, we introduce a formal bidding strategy model for specification of flexible and complex bidding strategies. The formal model is defined as a layered bidding strategy model (LBSM), which can be represented using notations borrowed from UML activity diagrams. To support real-time and efficient reasoning, the formal model is converted into a rulebased bidding strategy model (RBSM) specified in bidding strategy language (BSL) that can be directly executed by a reasoning module of a bidding agent. We present an algorithm for converting an LBSM to an RBSM, and an algorithm to drive the reasoning engine. Finally, we develop a prototype agent-based online auction system using JADE, and illustrate how flexible and complex bidding strategies can be precisely specified and efficiently executed. 1.

We present a general methodology to automate the search for equilibrium strategies in games derived from computational experimentation. Our approach interleaves empirical game-theoretic analysis with reinforcement learning. We apply this methodology to the classic Continuous Double Auction game, conducting the most comprehensive CDA strategic study published to date. Empirical game analysis confirms prior findings about the relative performance of known strategies. Reinforcement learning derives new bidding strategies from the empirical equilibrium environment. Iterative application of this approach yields strategies stronger than any other published CDA bidding policy, culminating in a new Nash equilibrium supported exclusively by our learned strategies.

We discuss CUECARD, the program that won the 2008 Computer Olympiad computational pool tournament. Beside addressing intrinsic interest in a complex competitive environment with unique features, our goal is to isolate the factors that contributed to the performance so that the lessons can be transferred to other, similar domains. Specifically, we distinguish among pure engineering factors (such as using a computer cluster), domainspecific factors (such as optimized break shots), and domain-independent factors (such as state clustering). Our conclusion is that each type of factor contributed to the performance of the program. 1

Experimental analysis of agent strategies in multiagent systems presents a tradeoff between granularity and statistical confidence. Collecting a large amount of data about each strategy profile improves confidence, but restricts the range of strategies and profiles that can be explored. We propose a flexible approach, where multiple game-theoretic formulations can be constructed to model the same underlying scenario (observation dataset). The prospect of incorrectly selecting an empirical model is termed generalization risk, and the generalization risk framework we describe provides a general criterion for empirical modeling choices, such as adoption of factored strategies or other structured representations of a game model. We propose a principled method of managing generalization risk to derive the optimal game-theoretic model for the observed data in a restricted class of models. Application to a large dataset generated from a trading agent scenario validates the method.

doi:10.1145/1378704.1378721 The most dramatic interaction between CS and GT may involve game-theory pragmatics.

We introduce the TAC Ad Auctions game (TAC/AA), a new game for the Trading Agent Competition. The Ad Auctions game investigates complex strategic issues found in real sponsored search auctions that are not captured in current analytical models. We provide an overview of TAC/AA, introducing its key features and design rationale. TAC/AA will debut in summer 2009, with the final tournament commencing in conjunction with the TADA-09 workshop.

Abstract not found

Online auctions have become extremely popular in recent years. Ability to predict winning bid prices accurately can help bidders to maximize their profit. This paper proposes a number of strategies and algorithms for performing such predictions for the first price sealed bid reverse auctions (FPSBRA). The Neural Networks (NN) and Genetic Programming (GP) learning techniques are used in the models. The algorithms are tested in the Trading Agent Competition Supply Chain Management (TAC SCM) game, where manufacture agents compete for customers ’ orders following the rules of the FPSBRA. Although all the proposed algorithms demonstrate the potential for predicting winning bid prices in competitive and dynamic environments, some of them perform more accurately than the others. 1.

Abstract not found